Makes sense as a pydantic model_validator.
We only support float and int dtypes right now. We don't support boolean, for instance. But the user can meddle with a raster internally to change the datatype.
We should have a validation error to protect against this happening.
It might be quite hard to do if numpy arrays can have their dtype modified in-place. Add some tests to try and understand that if it is a limitation.
Makes sense as a pydantic
model_validator.We only support float and int dtypes right now. We don't support boolean, for instance. But the user can meddle with a raster internally to change the datatype.
We should have a validation error to protect against this happening.
It might be quite hard to do if numpy arrays can have their dtype modified in-place. Add some tests to try and understand that if it is a limitation.